Sie sind auf Seite 1von 39

1

Quote is from John C. Maxwell, Ethics 101: What Every Leader Needs to Know, New York: Center Street, 2003, pp. 5051.

10

11

12

13

23

The following describes SELs recommended approach to relay testing and best practices: p g g g , , y g, 1. Perform comprehensive commissioning testing at the time of installation. Use thorough checklists, simulations, laboratory testing, and/or field checks to verify the performance of the protection system, including inputs, outputs, and settings. 2. Monitor the relay self-test alarm contact in real time via SCADA (supervisory control and data acquisition) or other monitoring system. If an alarm contact asserts, take immediate steps to repair, replace, or take corrective action for the alarmed relay. 3. Monitor potential relay failures not detected by self-tests. Specifically, these are logic inputs, contact outputs, and analog (voltage and current) inputs. Use continuous check of inputs (e.g., loss-of-potential logic) when available. If a secondary relay system is in place, compare the metering values between the primary and secondary systems. 4. 4 Analyze event reports to root cause and verify logic inputs and output contact operation. Use event reports as documentation to cause, operation validate correct operation of the protection system. If Steps 1 through 4 are followed, no periodic testing is required. Many users follow Steps 1 and 2 but do not perform Steps 3 and 4 consistently. In this case, perform periodic testing (e.g., once every ten years) on portions of the relay not tested by self-tests. This includes injecting known current and voltage signals to verify relay measuring accuracy, asserting inputs, and pulsing output contacts. This need not include re-verifying settings, plotting time-current curves or mho circles, etc. These characteristics are verified at commissioning and do not change. This testing may identify the small number of failures not detected by self-tests (e.g., using the p y ( g, g previous failure rates, testing every ten y , g y years would detect 0.05% 10 = 0.5% failures). ) This philosophy also applies to systems that do not operate frequently (e.g., bus or transformer differential protection). In some rare applications (installations without SCADA or communications), the self-test alarm is not monitored. This is not recommended. For these applications, the relays should be tested every one to five years, including the following: Check that the self-test alarm is not asserted. Inject known current and voltage signals to verify proper metering. Assert inputs, and pulse outputs. This need not include re-verifying settings, plotting time-current curves or mho circles, etc. These characteristics are verified at commissioning and do not change.

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39